Kevin Lomangino is the managing editor of HealthNewsReview.org. He tweets as @Klomangino.
“For heart health, it’s not just what you eat, but when.”
That was CBS’s confident headline about a scientific statement put out by the American Heart Association.
That statement, Meal Timing and Frequency: Implications for Cardiovascular Disease Prevention, summarizes the scientific evidence on issues like breakfast skipping, periodic fasting, and whether those types of eating patterns have any effect on the risk of cardiovascular disease.
The evidence is not strong, consisting mainly of epidemiological studies that assessed people’s eating habits and correlated them with risk of cardiovascular disease. Those types of studies, as we’ve pointed out many times, are not designed to support cause-and-effect conclusions.
There were also some clinical trials that assigned people to different eating patterns (e.g. eating breakfast vs. no breakfast) and tracked the impact on markers of cardiovascular risk – things like cholesterol levels, blood pressure, and blood sugar levels. But again, it’s unclear whether these surrogate outcomes will accurately reflect the impact on actual rates of disease.
In short, there was nothing definitive – no large randomized trial that assessed the impact of meal timing on outcomes that people care about. Nothing that would confidently allow us to say that eating breakfast actually lowered the rate of heart attacks and strokes. The studies that were reviewed had considerable flaws and limitations.
And yet the headlines were very assertive:
ReutersHealth: Skipping breakfast may be bad for your heart, doctors say
HealthDay: Timing of your meals might reduce heart risk
It should be noted that there were valuable cautions and caveats sprinkled throughout some of these stories (but not others). For example, HealthDay paraphrased the lead author of the AHA statement saying, “The evidence… is just not clear enough to make specific recommendations on breakfast.”
TIME helpfully explained, “The existing research isn’t strong enough yet to say that people who don’t normally eat breakfast should start—or that people who already do should expect to be heart-disease and diabetes-free for the rest of their lives.”
And some of these stories, like Reuters Health’s piece, did consult an outside expert for perspective on the research.
But those caveats are pushing back weakly against much stronger headline messages. And none of these stories got even close to what I thought was a deeper issue that reporters should have been raising: Why is the AHA issuing broad recommendations to the public based on evidence that’s so preliminary?
Ethan Weiss, MD, a cardiologist at University of California, San Francisco, wasn’t shy about voicing that concern. He took to Twitter to complain that the AHA had gone too far in issuing recommendations to the public — suggesting that a data summary targeted at researchers would have been more appropriate given the state of the science.
@PauliOhukainen why not just summarize the data and then say, “hey this is a super interesting area we should try to study carefully?”
— Ethan Weiss (@ethanjweiss) February 1, 2017
When I got him on the phone later in the day, Weiss described the issue to me this way:
“The problem is that the AHA is putting out a statement that’s going to be interpreted as being prescriptive, as in, ‘This is what you should do to be heart-healthy’ – when the reality, when you look at the summary, is that it says, ‘Basically we can’t say. This may or may not be important.’”
We’ve been here before, Weiss noted, when guidelines prematurely touted low-fat diets as an effective approach to heart disease prevention–inadvertently tarnishing “good” fats and possibly encouraging more consumption of refined carbohydrates.
“This is why we continue to get into trouble over and over again with nutrition,” he said. “We continue to make recommendations based on science that’s not solid.”
Yoni Freedhoff, MD, an assistant professor of family medicine at the University of Ottawa who’s also director of the Bariatric Medical Institute – a weight management clinic – agreed that the evidence available so far doesn’t support drawing any broad conclusions.
“I struggle with meal timing and importance studies because we all eat very different meals and lead hugely different lives,” said Freedhoff, who’s also an editorial contributor to HealthNewsReview.org. “Looking to breakfast specifically, the meal most hotly debated, can we really make conclusions about whether or not it’s helpful to satiety or the prevention of chronic disease if we’re lumping in bowls of Froot Loops with a couple of eggs and a slice of whole grain toast?”
Freedhoff added, “The gist, of course, is simply that we can’t comment on breakfast, lunch, or dinner if we’re not actually controlling for what people eat for breakfast, lunch or dinner. That would apply of course too to studies on meal frequency – if we’re not controlling for what people eat at particular meals, do we really have groups to compare?”
I’m not suggesting it’s easy to know when it’s appropriate to make recommendations about what people should and shouldn’t eat — or when news outlets should and shouldn’t cover those recommendations. Evidence in many areas of nutrition is notoriously weak, and that’s not going to change any time soon. So what do we base our recommendations on? Do we need a large randomized trial on broccoli before we recommend that people eat it?
Freedhoff thinks there’s sufficient evidence to make recommendations about broad-based eating patterns, for instance diets richer in fruits and vegetables being strongly correlated with decreased risks of various chronic non-communicable diseases. For less established areas of research, he said it’s reasonable to demand that health organizations and news outlets be more cautious in how they frame messages — especially when the science is as preliminary as it is with meal timing.
“I think the most important point is not drawing conclusions beyond those of the study and reflecting on study design prior to writing a headline, ” he said. “Again, breakfast stories are a great example given there is no definition for what constitutes breakfast and almost inevitably studies/headlines make broad sweeping conclusions.”